From: why the lucky stiff <yaml-core@wh...> - 2003-08-14 22:46:15
On Thursday 14 August 2003 01:35 pm, you wrote:
> You could let the user decide where the anchors should go.
> Some languages have reference counting, so users could
> use those decide on the anchors.
You're right, I should provide means for the user to override the anchori=
> (OK, someone might point into the graph from the outside,
> but you would typically only add a few redundant anchors.)
I wouldn't use reference counting. If I did, then in almost every case t=
top node would have an anchor.
> This is what the use of my C++ wrapper looks like, btw:
> (Alas, C++ doens't have introspection)
> .. definitions ...
> and then, to read a struct2:
> struct2 * pdg =3D yaml::Load<struct2>(fp);
Wow. This is cool!
You have much better ideas than I do for emitting. I wrote the current=20
emitter with very few ideas and I finished it just to finish it. But you=
sparking some very cool ideas.
Keep talking. Let me know where you go with this. (Oh, and post to the=20
yaml-core list as well. All this Syck e-mail we're exchanging is complet=
pertinent there. Plus, it provides some documentation for others who are=
quietly becoming Syckos.)
From: why the lucky stiff <yaml-core@wh...> - 2003-08-14 22:52:24
On Thursday 14 August 2003 01:38 pm, you wrote:
> No problem. I would still like to hear your views on the
> SYMID=3D=3D0 issue, though. Right now, I have this in my code:
Okay, then. SYMID of 0 is officially an invalid ID. I've checked in a n=
syck_add_sym() which starts symbol indexing with 1.
> Note the "if (!n->id)" wart.
You can remove the !n->id from your code now.
On Thu, Aug 14, 2003 at 04:34:13PM -0600, why the lucky stiff wrote:
> Keep talking. Let me know where you go with this.
Talk is cheap.
Here's what I have so far.
It's still incomplete but it should show where I'm heading.